The insurance relies on a principle of solidarityBut that is now undermined by the algorithms with which our profiles have been created.
Since these algorithms grow to be more demanding and precise, the premiums are increasingly personalized. This signifies that “high -risk profiles” are excluded from insurance systems as an entire, for the reason that costs grow to be unaffordable. Personalization has a certain legitimacy, but have to be compensated for with fair access to insurance.
First, nonetheless, it’s important to grasp that insurance relies on a basic paradox. On the one hand, its principles are based on a collective mechanism through which everyone contributes to their means and benefits of solidarity within the event of loss. On the opposite hand, technological advances, the growing wealth of knowledge and increasingly refined insurance mathematical methods are urging on increasingly individual pricing.
This tension is an increasingly demanding legal framework that prohibits any type of discrimination based on sensitive data, which sometimes correlates with relevant risk aspects.
As a professor of mathematics on the University of Quebec in Montreal (UQAM), I’m a co-author of the and writer of the recently published book . I would really like to analyze the challenge of reconciling the pooling based on solidarity, which is the insurance basis, with the hyper-segmented prices-grade-without policyholder enabled by big data.
Prize segmentation
Insurance corporations have used the classification as a pillar of their business model for a very long time: age, gender, work, geographical area, damage history, etc.
1662 English statisticians John Graunt published the mortality invoicesThe first statistical evaluation of the London death names. 1693, The English astronomer Edmund Halley developed the primary mortality tableWhat made it possible to calculate life expectancy at any age.
This work laid the foundations for differentiated pricing based on age and gender, which for a very long time remained the 2 fundamental criteria for segmentation in life and death insurance.
At the identical time after Great fire from London The first fire insurance contracts were published in 1666: corporations collected data on the variety of constructing materials and the urban density. In the 18th and nineteenth centuries, the tariffs were segmented in accordance with the proximity of neighboring buildings and the presence of fireplace fighting services, which led to the primary “high risk areas” and “low -risk areas”.
With the rise in the auto within the 1910s and Nineteen Twenties, the American insurers began to systematically record the variety of claims, the age and gender of the drivers. Several within the Nineteen Twenties “Classes” The prices were determined: young drivers, female drivers and experienced drivers who made it possible to find out the premiums in accordance with certain profiles.
Nowadays, current access to highly developed algorithms, tools for machine learning and a flood of knowledge: onboard telematics, connected devices, geolocalization, driving or lifestyle behavior and far more. For insurers, the refinery segmentation lets you invoice every policyholder “At their true risk”, “ Reduction of the results of cross -subsidizing good risks to poor risks and at the identical time improving general profitability.
However, the mutuality that is simply too precise reduces reciprocity and might make the insurance very expensive and even not accessible to certain high -risk segments. As a result, current balance today is searching for a subtle balance between recording the proper information to distinguish between profiles and the preservation of the living capability of the insured community.
The illusion of the profit of personalization
In Europe that FIDA proposal for financial data access (FIDA) would enable insurers to access the financial data of people. The aim is to refine the knowledge of expenses and repayment behavior. In this context, the promise of ultra-personalized pricing brings hopes for lower premiums, but additionally the fears before excessive profile and significant exclusions.
In view of this latest data industrial current, many shoppers see personalization as a win-win approach: If I higher use my budget, I’ll profit from a reduction. When my savings and repayment habits are classified as virtuous, my medical insurance premium drops. When my financial profile improves, my house insurance becomes cheaper.
This logic of “Pay-as-you-live” or “Pay-show you-drive” It is appealing: Individuals imagine that they’ve control over their insurance costs through their lifestyle decisions.
However, some points are highlighted.
-
The principle of mutuality will not be neutralized; Those who cannot take over probably the most virtuous behavior remain depending on the solidarity of others. Even if probably the most endangered people pay individually, those that are less in danger proceed to bear a share of costs due to the principle of mutuality.
-
Information asymmetry is reinforced since the insurer knows the statistics higher than the client. Personalized offers are sometimes based on correlations, sometimes weak, the meaning of which is unknown to the client.
-
Highly personalized products can force probably the most endangered or reverse risks to do without insurance as an entire and thus undermine the mutuality.
This also means whether it is expanded by accessing financial data. “Personalization” will not be necessarily synonymous with “Authorization” For the buyer.
The legal framework
The development of massive data within the insurance raises necessary ethical and legal questions: To what extent can sensitive variables be used to predict risks?
In France and the European Union, laws prohibits discrimination based on protected criteria reminiscent of ethnic origin, gender, sexual orientation, disability, religious beliefs and far more. In the Solvency II Directive (EU), insurers must use risk models who’re “transparent” And not discriminatory.
(Unplash)
In contrast to the European Union, which prohibits differentiated pricing based on protected criteria (gender, origin, disability), the Quebec model offers a more permissible framework. While the Quebec charta for human rights and freedoms also prohibit discrimination, it offers insurers from exemptions: if an element is statistically relevant, pricing for age, gender or family status relies.
This practice, which was only approved on the idea of the correlation, raises questions.
Ethics, social responsibility of the insurers
Apart from mere legal compliance, the moral practices and the social responsibility of insurers are increasingly examined by consumer associations and the media that report incidents with algorithmic discrimination and exercise status pressure.
In the past few years, insurers have had to lift awareness of how they will guarantee fair access to their products for in need of protection for population groups in need of protection without affecting the financial viability of their portfolios. To avoid exclusion, some revolutionary models offer “solidarity” formulas or limited rates.
The insurers face steadily increasing transparency requirements. In order to avoid the perception of arbitrariness, you will need to clearly explain your price criteria and make your calculation methods accessible. After all, you may have to integrate data protection and privacy into the design of your products (“Privacy in accordance with design”) to take care of trust.
Insurers who’re in a position to reconcile personalization, fairness and inclusion have gotten a benchmark for ethically -minded customers.
Bring solidarity and data into harmony
As we are able to see, the challenge is considerable.
It requires nothing lower than to recuperate the insurance mathematical precision with the values ​​of redistribution and solidarity which have underpinned the insurance occupation.
The way forward for insurance is simply decided by solving this voltage. There might be no pure price discrimination nor easy illusory personalization. Instead, the insurance industry must reconcile the 2 so that everybody might help make a contribution at their risk and to profit from the mutual uncertainty of life.
